Social media as societal microcosm: A decade of LGBT Twitter conversations in Singapore.

Background: Singapore occupies a curious societal grey-area: a digitally savvy country with a colonial-remnant law against homosexuality (penal code 377A), widely acknowledged as non-proactively enforced, existing to placate a conservative society; hotly contested for years and finally repealed in Parliament in 2022. Within a national context of state-upheld heteronormativity, yet with homosexuality not entirely condemned, Singapore occupies a liminal space where subtle resistance is carefully negotiated, especially in online spaces.

Objectives: This study investigated LGBT-adjacent discussions across social media over a decade (2011-2021) for salient topics, sentiment distribution, emotional intensity frames and nuanced topics.

Methods: Historical Twitter data containing LGBT keywords (N = 15,659) were collected and analyzed using bi-term topic modelling, sentiment score modelling, and emotional intensity modelling. Qualitative thematic analysis was conducted on highest-scoring emotion tiers.

Results: Sentiment was distributed over a range: Very Positive (6%), Positive (33%), Neutral (11%), Negative (42%), Very Negative (7%). Predominant emotions were Joy (39%), Anger (32%), Sadness (11%), Fear (11%). Of themes from highest-scoring emotional-intensity tweets, Anger included: 'gay' used derogatorily; heated debates over ideological-often religious-differences; dissent within the community, condemning exclusionary views. Fear and Sadness included distress over violence (mass shootings, harassment, bullying); lack of acceptance (criminalization, protests over local pride event 'Pinkdot'; lack of familial support). Joy stemmed from the celebration of pride month.

Conclusions: Our findings highlight nuanced emotional intensities, profiles undertones of LGBT dissent and support, fractured along a schism of differing views and contrasting opinions-a societal microcosm of a divisive topic. Practically, this presents a decade-long barometer of dominant trigger points that may help facilitate conversations on the affective concerns of the local population.